Stein-Rule Estimation in Mixed Regression Models
Vorname Shalabh, Alan T. K. Wan · Biometrical Journal · 2000
This paper considers a Stein-rule mixed regression estimator for estimating a normal linear regression model in the presence of stochastic linear constraints. We derive the small disturbance asymptotic bias and risk of the proposed estimator, and analytically compare its risk with other related estimators. A Monte-Carlo experiment investigates the empirical risk performance of the proposed estimator.